NAG Fortran Library Routine Document D02MVF.1

Transcription

1 D02 Ordinary Differential Equations NAG Fortran Library Routine Document Note: before using this routine, please read the Users Note for your implementation to check the interpretation of bold italicised terms and other implementation-dependent details. 1 Purpose is a setup routine which must be must be called prior to an integrator in sub-chapter D02M/N, if the DASSL implementation of Backward Differentiation Formulae (BDF) is to be used. 2 Specification SUBROUTINE (NEQMAX, NY2DIM, MAXORD, CON, TCRIT, HMIN, HMAX, H0, 1 MAXSTP, MXHNIL, NORM, RWORK, IFAIL) INTEGER NEQMAX, NY2DIM, MAXORD, MAXSTP, MXHNIL, IFAIL double precision CON(3), TCRIT, HMIN, HMAX, H0, RWORK(50+4NEQMAX) CHARACTER1 NORM 3 Description An integrator setup routine must be called before the call to any integrator in this sub-chapter. This setup routine makes the choice of the DASSL integrator and permits you to define options appropriate to this choice. 4 References See the D02M/N Sub-chapter Introduction. 5 Parameters 1: NEQMAX INTEGER Input On entry: a bound on the maximum number of differential equations to be solved. Constraint: NEQMAX 1. 2: NY2DIM INTEGER Input On entry: the second dimension of the array YSAVE that will be supplied to the integrator, as declared in the (sub)program from which the integrator is called (e.g., see D02NBF). Constraint: NY2DIM MAXORD þ 3. 3: MAXORD INTEGER Input On entry: the maximum order to be used for the BDF method. If MAXORD ¼ 0 then MAXORD ¼ 5 is assumed. Constraint: 0 MAXORD 5. 4: CONð3Þ double precision array Input/Output On entry: values to be used to control step size choice during integration. If any CONðiÞ ¼0:0 on entry, it is replaced by its default value described below. In most cases this is the recommended setting. CONð1Þ, CONð2Þ, and CONð3Þ are factors used to bound step size changes. If the current step size h fails, then the modulus of the next step size is bounded by CONðlÞjj. h The default value of.1

2 NAG Fortran Library Manual CONð1Þ is 2:0. Note that the new step size may be used with a method of different order to the failed step. If the initial step size is h, then the modulus of the step size on the second step is bounded by CONð3Þjj. h At any other stage in the integration, if the current step size is h, then the modulus of the next step size is bounded by CONð2Þjj. h The default values are 10:0 for CONð2Þ and 1000:0 for CONð3Þ. Constraints: 0:0 < CONð1Þ < CONð2Þ < CONð3Þ; CONð2Þ > 1:0; CONð3Þ > 1:0. These constraints must be satisfied after any zero values have been replaced by default values. On exit: the values actually to be used by the integration routine. 5: TCRIT double precision Input On entry: a point beyond which integration must not be attempted. The use of TCRIT is described under the parameter ITASK in the specification for the integrator (e.g., see D02NBF). A value, 0:0 say, must be specified even if ITASK subsequently specifies that TCRIT will not be used. 6: HMIN double precision Input On entry: the minimum absolute step size to be allowed. Set HMIN ¼ 0:0 if this option is not required. 7: HMAX double precision Input On entry: the maximum absolute step size to be allowed. Set HMAX ¼ 0:0 if this option is not required. 8: H0 double precision Input On entry: the step size to be attempted on the first step. Set H0 ¼ 0:0 if the initial step size is to be calculated internally. 9: MAXSTP INTEGER Input On entry: the maximum number of steps to be attempted during one call to the integrator after which it will return with IFAIL ¼ 2 (e.g., see D02NBF). Set MAXSTP ¼ 0 if no limit is to be imposed. 10: MXHNIL INTEGER Input On entry: the maximum number of warnings printed (if ITRACE 0, e.g., see D02NBF) per problem when t þ h ¼ t on a step (h ¼ current step size). If MXHNIL 0, a default value of 10 is assumed. 11: NORM CHARACTER1 Input On entry: indicates the type of norm to be used. NORM ¼ M Maximum norm. NORM ¼ A Averaged L2 norm. NORM ¼ D Is the same as A..2

3 D02 Ordinary Differential Equations If vnorm denotes the norm of the vector v of length n, then for the averaged L2 norm vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi u X n 1 v 2 vnormb ¼ t i, while for the maximum norm n i¼1 vnorm ¼ max v i w i 1in w i If you wish to weight the maximum norm or the L2 norm, then RTOL and ATOL should be scaled appropriately on input to the integrator (see under ITOL in the specification of the integrator for the formulation of the weight vector w i from RTOL and ATOL, e.g., see D02NBF). Only the first character to the actual parameter NORM is passed to ; hence it is permissible for the actual argument to be more descriptive, e.g., Maximum, Average L2 or Default in a call to. Constraint: NORM ¼ M, A or D. 12: RWORKð50 þ 4 NEQMAXÞ double precision array Communication Array This must be the same workspace array as the array RWORK supplied to the integrator. It is used to pass information from the setup routine to the integrator and therefore the contents of this array must not be changed before calling the integrator. 13: IFAIL INTEGER Input/Output On entry: IFAIL must be set to 0, 1 or 1. If you are unfamiliar with this parameter you should refer to Chapter P01 for details. On exit: IFAIL ¼ 0 unless the routine detects an error (see Section 6). For environments where it might be inappropriate to halt program execution when an error is detected, the value 1 or 1 is recommended. If the output of error messages is undesirable, then the value 1 is recommended. Otherwise, if you are not familiar with this parameter the recommended value is 0. When the value 1 or 1 is used it is essential to test the value of IFAIL on exit. 6 Error Indicators and Warnings If on entry IFAIL ¼ 0or 1, explanatory error messages are output on the current error message unit (as defined by X04AAF). Errors or warnings detected by the routine: IFAIL ¼ 1 On entry, NEQMAX < 1, or NY2DIM < MAXORD þ 3, or MAXORD < 0, or MAXORD > 5, or invalid value for element of the array CON, or NORM 6¼ M, A or D. 7 Accuracy Not applicable. 8 Further Comments None...3

4 NAG Fortran Library Manual 9 Example We solve the plane pendulum problem defined by the equations: x 0 ¼ u y 0 ¼ v u 0 ¼ x v 0 ¼ y 1 x 2 þ y 2 ¼ 1 The additional algebraic constraint xu þ yv ¼ 0 can be derived, and after appropriate substitution and manipulation to avoid a singular Jacobian we solve the equations: y 0 1 ¼ y 3 y 6 y 1 y 0 2 ¼ y 4 y 6 y 2 y 0 3 ¼ y 5 y 1 y 0 4 ¼ y 5 y ¼ y 1 y 3 þ y 2 y 4 0 ¼ y 2 1 þ y with given initial conditions and derivatives. 9.1 Program Text Example Program Text Mark 14 Release. NAG Copyright Parameters.. INTEGER NOUT (NOUT=6) INTEGER NEQ, NEQMAX, NRW, NINF, NWKJAC, MAXORD, NY2DIM, + MAXSTP, MXHNIL (NEQ=6,NEQMAX=NEQ,NRW=50+4NEQMAX,NINF=23, + NWKJAC=NEQMAX(NEQMAX+1),MAXORD=5, + NY2DIM=MAXORD+3,MAXSTP=5000,MXHNIL=5) DOUBLE PRECISION H0, HMAX, HMIN (H0=0.0D0,HMAX=0.0D0,HMIN=1.0D-10).. Local Scalars.. DOUBLE PRECISION DUM, PI, T, TCRIT, TOUT INTEGER I, IFAIL, ITASK, ITOL, ITRACE, MAXOD1.. Local Arrays.. DOUBLE PRECISION ATOL(NEQMAX), CONST(3), RTOL(NEQMAX), RWORK(NRW), + WKJAC(NWKJAC), Y(NEQMAX), YDOT(NEQMAX), + YSAVE(NEQMAX,NY2DIM) INTEGER INFORM(NINF) LOGICAL LDERIV(2).. External Functions.. DOUBLE PRECISION X01AAF EXTERNAL X01AAF.. External Subroutines.. EXTERNAL, D02NBY, D02NGF, D02NSF, DAEJAC, DAERES, + X04ABF.. Executable Statements.. WRITE (NOUT,) Example Program Results CALL X04ABF(1,NOUT) DO 20 I = 1, 3 CONST(I) = 0.0D0 20 CONTINUE ITRACE = 0 PI = X01AAF(DUM) RTOL(1) = 1.0D-3 ATOL(1) = 1.0D-6 WRITE (NOUT,) WRITE (NOUT,99999) Pendulum problem with relative tolerance =, + RTOL(1) WRITE (NOUT,99999) and absolute tolerance =, + ATOL(1) ITOL = 1 T = 0.0D0.4

5 D02 Ordinary Differential Equations TOUT = PI Set initial values and derivatives Y(1) = 1.D0 Y(2) = 0.0D0 Y(3) = 0.D0 Y(4) = 0.D0 Y(5) = 0.0D0 Y(6) = 0.0D0 YDOT(1) = Y(3) - Y(6)Y(1) YDOT(2) = Y(4) - Y(6)Y(2) YDOT(3) = -Y(5)Y(1) YDOT(4) = -Y(5)Y(2) - 1.D0 YDOT(5) = -3.D0Y(4) YDOT(6) = 0.0D0 WRITE (NOUT,) WRITE (NOUT,) + t y1 y2 y3 y4 y5 y6 WRITE (NOUT,99998) T, (Y(I),I=1,6) ITASK = 4 TCRIT = TOUT IFAIL = 0 MAXOD1 = 0 CALL (NEQMAX,NY2DIM,MAXOD1,CONST,TCRIT,HMIN,HMAX,H0,MAXSTP, + MXHNIL, AVERAGE-L2,RWORK,IFAIL) CALL D02NSF(NEQ,NEQMAX, ANALYTIC,NWKJAC,RWORK,IFAIL) IFAIL = -1 LDERIV(1) =.TRUE. LDERIV(2) =.TRUE. CALL D02NGF(NEQ,NEQMAX,T,TOUT,Y,YDOT,RWORK,RTOL,ATOL,ITOL,INFORM, + DAERES,YSAVE,NY2DIM,DAEJAC,WKJAC,NWKJAC,D02NBY,LDERIV, + ITASK,ITRACE,IFAIL) IF (IFAIL.NE.0) THEN WRITE (NOUT,99997) Integration terminated with IFAIL =, + IFAIL, at T =, T ELSE WRITE (NOUT,99998) T, (Y(I),I=1,NEQ) END IF STOP FORMAT (1X,A,1P,E7.1) FORMAT (1X,F6.4,3X,6(F8.4)) FORMAT (1X,A,I2,A,F6.4) END SUBROUTINE DAERES(NEQ,T,Y,YDOT,R,IRES).. Scalar Arguments.. DOUBLE PRECISION T INTEGER IRES, NEQ.. Array Arguments.. DOUBLE PRECISION R(NEQ), Y(NEQ), YDOT(NEQ).. Executable Statements.. IF (IRES.EQ.-1) THEN R(1) = -YDOT(1) R(2) = -YDOT(2) R(3) = -YDOT(3) R(4) = -YDOT(4) R(5) = 0.0D0 R(6) = 0.0D0 ELSE R(1) = Y(3) - Y(6)Y(1) - YDOT(1) R(2) = Y(4) - Y(6)Y(2) - YDOT(2) R(3) = -Y(5)Y(1) - YDOT(3) R(4) = -Y(5)Y(2) - 1.D0 - YDOT(4) R(5) = Y(1)Y(3) + Y(2)Y(4) R(6) = Y(1)2 + Y(2)2-1.0D0 END IF RETURN.5

6 NAG Fortran Library Manual END SUBROUTINE DAEJAC(NEQ,T,Y,YDOT,H,D,P).. Parameters.. DOUBLE PRECISION ONE, TWO (ONE=1.0D0,TWO=2.0D0).. Scalar Arguments.. DOUBLE PRECISION D, H, T INTEGER NEQ.. Array Arguments.. DOUBLE PRECISION P(NEQ,NEQ), Y(NEQ), YDOT(NEQ).. Local Scalars.. DOUBLE PRECISION HXD.. Executable Statements.. HXD = HD P(1,1) = (ONE+HXDY(6)) P(1,3) = -HXD P(1,6) = HXDY(1) P(2,2) = (ONE+HXDY(6)) P(2,4) = -HXD P(2,6) = HXDY(2) P(3,1) = HXDY(5) P(3,3) = ONE P(3,5) = HXDY(1) P(4,2) = HXDY(5) P(4,4) = ONE P(4,5) = HXDY(2) P(5,1) = -HXDY(3) P(5,2) = -HXDY(4) P(5,3) = -HXDY(1) P(5,4) = -HXDY(2) P(6,1) = -TWOHXDY(1) P(6,2) = -TWOHXDY(2) RETURN END 9.2 Program Data None. 9.3 Program Results Example Program Results Pendulum problem with relative tolerance = 1.0E-03 and absolute tolerance = 1.0E-06 t y1 y2 y3 y4 y5 y (last)